JP2005035533A - Travel load information learning system and travel load information learning method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To learn travel load information as a basic data instead of prestoring a basic data of the travel load leaning information in a CD-ROM or DVD-ROM. <P>SOLUTION: When a vehicle travels on a road, a current position of the traveling vehicle is found by utilizing road map information used in a navigation system. Travel load information of the road of the current position is estimated, and the traveling load information associated with the current position is learned and stored in a map DB 4c of a navigation ECU 4. In this way, travel load information is learned while the vehicle travels. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a traveling load information learning method, a traveling load information learning system, and a vehicle driving force control system using the traveling load information learning method for learning traveling load information such as a road surface gradient, turning curvature, and inclination. .

  2. Description of the Related Art Conventionally, there is a vehicle driving force control device that corrects an engine control amount based on a slope of a road surface on which the vehicle is traveling so that the driver does not feel deficient in driving force or poor reaction (for example, Patent Document 1). reference). Specifically, the vehicle driving force control device uses road map information held in a CD-ROM or DVD-ROM of a navigation system, and gradient information associated with the road map information, and uses the road map information. After calculating the gradient difference between the current position of the host vehicle and the estimated position where the host vehicle will be traveling several seconds later, the engine control amount is corrected based on the calculation result.

Further, this vehicle driving force control device stores the vehicle speed information and driving state associated with the road map information in the navigation ECU every time the vehicle travels, and uses it together with auxiliary information such as gradient information stored in the navigation ECU. The engine control amount is corrected.
Japanese Patent No. 3203976

  However, the conventional vehicular driving force control device holds in advance the basic data of the parameters for correcting the engine control amount, such as gradient information, in the CD-ROM or DVD-ROM for the navigation system. Is assumed. And the calculation of correction | amendment of engine control amount is performed using the basic data. For this reason, it is necessary to create basic data in advance, and there is also a problem that data memory such as CD-ROM and DVD-ROM increases because the basic data must be stored.

  Also, the shape and slope of the road are changed year by year, and unless the CD-ROM or DVD-ROM is replaced every time, the engine control amount corresponding to the change cannot be corrected.

  The present invention has been made in view of the above problems, and allows basic road load information to be learned as basic data, instead of storing basic data of the road load information in a CD-ROM, DVD-ROM or the like in advance. It is an object to provide a traveling load information learning method, a traveling load information learning system, and a vehicle driving force control system using the same.

  In order to achieve the above object, according to the first aspect of the present invention, current position detecting means (4, 4c) for obtaining a current position where the vehicle is traveling based on road map information when the vehicle travels on the road; Travel load information estimation means (4, 4a, 4b, 6, 6a, 6b) for estimating road travel load information at the current position, and storage means for learning and storing travel load information in association with the current position ( 4, 4c).

  According to such a configuration, when the vehicle travels, road load information on the road can be estimated, and the travel load information can be learned and stored. As a result, it is possible to learn road surface traveling load information as basic data instead of storing basic data of traveling load information in a CD-ROM or DVD-ROM in advance.

  For example, as shown in claim 2, the road load information includes road gradient and curvature.

  In the third aspect of the invention, the storage unit collects the travel load information estimated by the travel load information estimation unit and stores it according to at least one of the travel amount in the predetermined area, the reliability of the estimated information, and the learning level. The power value is obtained and stored as learning content of the traveling load information.

  Thus, the storage means collects the travel load information estimated by the travel load information estimation means, obtains a value to be stored according to at least one of the travel amount of the predetermined area, the reliability of the estimated information, and the learning degree, It is stored as learning content of travel load information. For example, more accurate traveling load can be obtained by performing weighted addition according to at least one of the time when the traveling load information is estimated, the parameter used for estimating the traveling load information, and the state of the traveling load information estimating means. It becomes possible to ask for information.

  In the invention according to claim 4, the storage means stores traveling load information together with ID data indicating a predetermined space area corresponding to the current position, and the traveling load information estimation means travels in the predetermined space area. Instantaneous information is collected, and an average value in the size of the predetermined space area is estimated as the traveling load information to be stored in the storage means.

  In this way, instantaneous information when traveling in a predetermined space area is collected, and the average value of the size of the predetermined space area, for example, distance, time, and area is stored in the storage means. It can be estimated as travel load information to be made. For example, the storage means stores travel load information together with ID data indicating a segment or node corresponding to the current position, and the travel load information estimation means is provided every time a predetermined distance is advanced between the segments or nodes. By accumulating the obtained instantaneous information and dividing the accumulated value of the instantaneous information by the moving distance of the vehicle, the running load information can be estimated by obtaining an average value between segments or nodes.

  In this case, as shown in claim 5, when accumulating instantaneous information, it corresponds to at least one of cases where the vehicle is at a very low speed, when a slip occurs, or when there is a malfunction in the travel load information estimating means In that case, the average value may be obtained by removing the instantaneous information in that case from the accumulation. In this way, more accurate travel load information can be estimated.

  In the invention according to claim 6, the storage means stores the number of times of learning of the travel load information, and the travel load information estimated this time by the travel load information estimation means and the travel load previously stored in the storage means. From the information, an average value of travel load information corresponding to the number of learning is obtained, and the average value is stored as new travel load information.

  In this way, when there is travel load information stored previously, the travel according to the number of times of learning is calculated from the travel load information estimated this time and the travel load information stored previously in the storage means. By obtaining the average value of the load information, it is possible to obtain the traveling load information. Thus, more accurate traveling load information can be obtained by using traveling load information obtained in the past.

  In such a travel load information learning system according to any one of claims 1 to 6, as shown in claim 7, data relating to travel load information stored in the storage means from the output means (4f). Can be output.

  For this reason, it becomes possible to perform vehicle control according to traveling load information by inputting the data regarding such traveling load information into the control part which performs vehicle control, for example, vehicle driving force control.

  The invention according to claim 8 includes a branch detection means (4) for obtaining a branch of the road on which the vehicle is traveling based on the road map information, and detected by the branch detection means together with the travel load information from the output means. It is characterized by outputting information about branches.

  Thus, if the information regarding a branch is output with driving load information, the control part which performs the vehicle control which input it can perform the vehicle control corresponding to a branch.

  The invention according to claim 9 further comprises estimated section detecting means (4) for detecting a section where the travel load information is estimated and a section where the travel load information is not estimated based on the road map information. Along with the information, information relating to a place where the travel load information detected by the estimation section detecting means is not estimated is output.

  In this way, by outputting information related to the place where the travel load information is not estimated, the control unit that executes vehicle control that has input the vehicle load information corresponding to the place where the travel load information is not estimated. Control can be executed.

  In the invention according to claim 10, it is partially erased according to the storage amount or storage allowable amount of the storage means, and erased according to at least one of the learning time, the learning level, the current location or the spatial relationship with the predetermined location. It is characterized by what is to be decided.

  As described above, the object to be erased can be determined based on the learning time, the degree of learning, the current location, or the spatial relationship with a predetermined location (for example, home).

  In the invention described in claim 11, the storage means forms a structure in which the travel load information is associated with the travel point, and is divided into at least management information and travel environment information, the learning degree is management information, and the load information is travel environment. It is characterized by registering as information.

  As described above, the travel load information can be a structure associated with the travel point, divided into at least management information and travel environment information, the learning degree can be registered as management information, and the load information can be registered as travel environment information. In this way, if the configuration is stratified according to the contents to be stored, it is easy to organize at the time of extraction, and it is possible to cope with the presence / absence of functions and addition / deletion of applications using general-purpose access software.

  The invention described in claim 12 is a vehicle driving force control system having the travel load information learning system according to any one of claims 1 to 9, and based on the travel load information and road map information stored in the storage means, A driving force control unit (6) that predicts a driving load of a road on which the vehicle will travel and controls the driving force of the vehicle according to the prediction result is provided.

  As described above, the road load information stored in the road load information learning system and the road map information used in the navigation system or the like are used to predict the road load on which the vehicle will travel, and according to the prediction result. Thus, the driving force of the vehicle can be controlled.

  In this case, as shown in claim 13, the driving force control unit can adjust the gear stage in the transmission (7) of the vehicle according to the prediction result. In this way, the range of driving force that can be operated according to the driver's will is compared with the case where the engine control amount is corrected because the vehicle is adapted to the road load on which the vehicle will travel by shifting the transmission. Can be optimized.

  In the invention according to claim 14, when the vehicle travels on the road, the current position where the vehicle is traveling is obtained based on the road map information, and after the road load information on the road at the obtained current position is estimated, It is characterized in that travel load information is learned and stored in association with the position. This claim corresponds to the form in which claim 1 is changed to the traveling load information learning method, and the same effect as that of claim 1 can be obtained.

  In addition, the code | symbol in the bracket | parenthesis of each said means shows the correspondence with the specific means as described in embodiment mentioned later.

(First embodiment)
A vehicle driving force control system including a traveling load information learning system according to a first embodiment of the present invention will be described.

  FIG. 1 shows a schematic configuration of a vehicle drive control device. As shown in this figure, the vehicle driving force control system includes various sensor groups 1a to 1d, 2a to 2d, 3a and 3b, a navigation ECU 4, an engine ECU 6 for controlling the engine 5, and a transmission (hereinafter referred to as a vehicle). The T / M-ECU 8 and the brake ECU 9 are configured to control 7).

  Various sensor groups 1a to 1d, 2a to 2d are a wheel speed sensor 1a to 1d, an engine speed sensor 2a, a throttle opening sensor 2b, an intake air temperature sensor 2c, an accelerator pedal sensor 2d, a yaw rate sensor (gyro sensor) 3a, and The shift position sensor 3b is provided.

  The wheel speed sensors 1a to 1d are provided for each wheel FR, FL, RR, RL provided in the vehicle. The wheel speed signals of the corresponding wheels FR, FL, RR, RL are output from the respective wheel speed sensors 1a to 1d. As is well known, each wheel speed signal is used for calculation of wheel speed, calculation of vehicle speed (body speed), etc., in each wheel FR, FL, RR, RL.

  The engine speed sensor 2a is for obtaining the rotational speed of the engine 5 which is a power source for generating power. The throttle opening sensor 2b is for detecting the throttle opening of the engine 5 and obtaining the intake air amount of the engine 5 based on the obtained throttle opening. The intake air temperature sensor 2 c is for detecting the intake air temperature of the engine 5. The accelerator sensor 2d detects a driver's driving force request according to an operation of an accelerator pedal (not shown).

  The yaw rate sensor 3a detects a yaw rate (yaw angular velocity) of the vehicle, and is generally provided in a navigation system.

  The navigation ECU 4 constitutes a part of the navigation system and has a function of executing general navigation control using road map information stored in a storage medium such as a CD-ROM or DVD-ROM in the navigation system. In addition, it has a function of learning road gradient and curvature. The navigation ECU 4 is provided with a storage unit (not shown), and the learned road gradient and curvature are stored in the storage unit as a learning gradient and a learning curvature. These learning gradients and learning curvatures are used as parameters for various calculations executed by the engine ECU 6. The navigation ECU 4 calculates an instantaneous curvature (hereinafter referred to as an instantaneous curvature) that is calculated every time the vehicle travels on the road, and calculates a learning gradient and a learning curvature.

  The engine ECU 6 corresponds to power adjustment means, and performs engine control according to the amount of operation of the accelerator pedal based on the detection signal from the accelerator pedal sensor 2d. The engine ECU 6 outputs an engine control signal indicating a power command value, and engine control is performed based on the engine control signal. The engine ECU 6 calculates an instantaneous gradient (hereinafter referred to as an instantaneous gradient) that is calculated every time the vehicle travels on the road, and according to the driving environment based on the learning gradient and the learning curvature obtained by the navigation ECU 4. Find the load fluctuation.

  The T / M-ECU 8 performs T / M 7 gear switching (shift) according to the position of a shift lever (not shown), and outputs a T / M control signal to the T / M-ECU 8 to output T / M-ECU 8. / M-ECU8 is shifting. Further, when the shift lever is in the D (driving) range, an optimum gear stage corresponding to the driver request torque received from the engine ECU 6 and the driving environment load fluctuation is obtained, and T / The gear stage of M7 is controlled.

  The brake ECU 9 executes braking force control, and performs brake control according to an operation amount of a brake pedal (not shown). The brake ECU 9 receives wheel speed signals from the wheel speed sensors 1a to 1d, calculates wheel speeds and vehicle speeds based on the wheel speed signals, and performs various brake controls based on these. In this embodiment, the brake ECU 9 is used to transmit the vehicle speed calculated based on the wheel speed signals from the wheel speed sensors 1a to 1d to the navigation ECU 4 and the engine ECU 6.

  Since the T / M-ECU 8 receives the rotation speed signal of the output shaft of the T / M 7, even if the vehicle speed is not transmitted from the brake ECU 9 due to a failure or the like, the rotation of the output shaft of the T / M 7 It is possible to independently determine the vehicle speed based on the number signal.

  Next, the configuration and operation of the navigation ECU 4 and the engine ECU 6 provided in the vehicle driving force control device will be described in detail.

  The navigation ECU 4 has a block configuration shown in FIG. As shown in this figure, the navigation ECU 4 includes a yaw rate estimation unit 4a and a curvature estimation unit 4b. The yaw rate estimation unit 4a estimates the yaw rate based on a detection signal from a yaw rate sensor 3a that is generally provided in the navigation system. Moreover, the curvature estimation part 4b estimates the curvature of a road surface based on the estimated yaw rate and vehicle speed. Specifically, when the yaw rate is γ [rad / sec], the curvature is ρ [1 / m], and the vehicle speed is v [m / sec], these satisfy the relationship of γ = ρ × v. Based on the vehicle speed and yaw rate at that time, the curvature is obtained.

  As described above, the navigation ECU 4 calculates the curvature of the road surface on which the vehicle travels during each cycle, that is, the instantaneous curvature, for each predetermined calculation cycle.

  On the other hand, the engine ECU 6 has a block configuration shown in FIG. As shown in this figure, the engine ECU 6 includes a drive torque estimating unit 6a and a gradient estimating unit 6b. The drive torque estimating unit 6a receives signals from the engine speed sensor 2a, the throttle opening sensor 2b, the intake air temperature sensor 2c, and the accelerator pedal sensor 2d, and estimates the drive torque based on these signals. This drive torque is estimated by a detection signal from the engine speed sensor 2a, an intake air amount obtained based on a detection signal from the throttle sensor 2b, a detection signal from the intake air temperature sensor 2c, and the engine ECU 6 itself. This is performed by a known method based on the fuel injection amount and the plug ignition timing which are calculated and adjusted.

  The gradient estimation unit 4b estimates the gradient of the road surface on which the vehicle has traveled based on the estimated driving torque. The estimation of the road gradient is executed based on the driving torque, the gear ratio of the gear at T / M7, and the like. The estimation of the road surface gradient is performed, for example, in the same manner as in the method disclosed in Japanese Patent Laid-Open No. 8-53024, and thus the description thereof is omitted here.

  As described above, the engine ECU 6 calculates the gradient of the road surface on which the vehicle travels during each cycle, that is, the instantaneous gradient, for each predetermined calculation cycle. This instantaneous gradient is sent to the navigation ECU 4 through the in-vehicle LAN.

  Further, the navigation ECU 4 further has a block configuration shown in FIG. As shown in this figure, the navigation ECU 4 includes a map database (hereinafter referred to as a map DB) 4c corresponding to a storage unit. The navigation ECU 4 learns the instantaneous gradient sent from the engine ECU 6 and the instantaneous curvature calculated by the navigation ECU 4 as travel load information and stores it in the map DB 4c.

  Specifically, the navigation ECU 4 corrects the calculation error of the relative position of the vehicle estimated from the result calculated based on the signals obtained from the gyro sensor and the vehicle speed sensor by the absolute position obtained from the GPS signal. Estimate the exact current position. Then, the map matching technique is used to collate with the position coordinate sequence recorded in the map DB 4c, estimate the road on which the vehicle is traveling, and extract information on the traveling road.

  Since the current position calculation method and map matching technique are well known, description thereof is omitted here.

  Then, the navigation ECU 4 stores the instantaneous gradient and the instantaneous curvature in association with the current position. For example, in a CD-ROM or DVD-ROM used for a navigation system in which road map information is stored, nodes indicating arbitrary points on the road and segments (links) connecting the nodes are defined. Basic data on roads is stored for each segment. An ID address is determined for each node and segment so that each node and segment can be identified. For this reason, the navigation ECU 4 stores the instantaneous gradient, the instantaneous curvature, and the date and time when they are obtained in the memory corresponding to the ID address of the node or segment in the storage unit provided in the navigation ECU 4 or together with the ID address in the storage unit. Memorize etc. The instantaneous gradient and the instantaneous curvature stored in this way correspond to the learning gradient and the learning curvature.

  FIG. 5 shows the relationship between the contents stored in a CD-ROM or DVD-ROM used in the navigation system and the contents stored in the storage unit provided in the navigation ECU 4.

  As shown in this figure, for example, in a CD-ROM or DVD-ROM used in a navigation system, an ID address is determined for each segment, and a segment start point ID, end point ID, road, etc. are used as basic data. The type is stored. For this reason, the storage unit in the navigation ECU 4 stores the instantaneous gradient, the instantaneous curvature, the horizontal inclination (hereinafter referred to as “kant”), the number of learnings, and the like together with the ID address.

  At this time, the memory configuration in the storage unit is a structure in which memories are stratified according to contents to be stored such as experience management data, environmental data, vehicle data, internal or device level, and the like. And it is divided into at least management information and driving environment information, the learning degree is registered as management information, and the load information is registered as driving environment information. For example, the memory for storing the experience management data stores the number of times of learning, the number of passages of nodes or segments, the latest passage date and time, and the like. A learning gradient, a learning curvature, a learning cant, a learning trajectory coordinate, and the like are stored in the memory that stores environment data. The memory for storing the vehicle data stores control conditions / parameters, traveling vehicle speed, and the like. The memory for storing the internal or device level stores the position jumping experience, the experience of being unable to capture the satellite, taking a course along the road, and the like.

  In this way, if the memory configuration is such that the memory is stratified according to the contents to be stored, it is easy to organize at the time of extraction, and it is possible to deal with the presence or absence of functions and addition / deletion of applications using general-purpose access software Become.

  When these are stored, the storage unit provided in the navigation ECU 4 is stored in a memory area that cannot be accessed by a general user, for example, a part such as a system area and not a disk cache area. Is preferred. For example, a partition other than a partition accessible by a general user, a folder whose access is prohibited, and a flash memory may be mentioned.

  The following is executed as a pre-process for storing in the memory. In other words, it is determined which node the current driving point is between, and every time a predetermined distance (for example, 1 m) is passed from the point of passing through the node to the next node, the instantaneous gradient, the instantaneous curvature, and the moving distance are respectively determined. Accumulate. Then, when the next node is reached, the cumulative value is divided by the movement distance to obtain the average value for that section. Or, every time a predetermined time (for example, 1 second) elapses, the instantaneous information and the elapsed time are accumulated, and when the delimiter is exceeded, the instantaneous information accumulated value is divided by the elapsed time accumulated value to obtain an average value for that section. And As described above, the average value of the travel amount such as the distance and time when traveling in the predetermined area is stored in the memory corresponding to the ID address of the segment, for example.

  However, when an extremely low speed or slip occurs, gradient estimation is performed, and when there is a problem in the engine ECU 6 that is an information transmission source that transmits data related to the instantaneous gradient to the navigation ECU 4, or the current position is confident in detection. When it is low, the weighting factor W (W is a value from 0 to 1) is determined according to the speed, slip state, failure occurrence status, position detection confidence level, and the average processing is obtained by the following equation: It becomes possible to improve the reliability of the learning gradient and the learning extreme rate.

  Since the learning gradient and the learning curvature are obtained every time the vehicle travels on the road, even if learning is interrupted as appropriate in the above case, learning is usually performed once. In this case, the road has been traveled again, and the previously calculated learning gradient and learning curvature may already be stored. In such a case, an average value with the previously stored learning gradient and learning curvature is calculated, and the average value is stored. Specifically, assuming that the learning data stored before is C, the learning data calculated this time is c, and the number of learnings is n, the average value C ′ is obtained by the following equation, and a new learning gradient is obtained. , Stored as learning curvature.

なお、ここでいう学習回数は、ナビゲーションＥＣＵ４に学習回数を示すカウンタを備えておき、瞬時勾配が求められるたびにカウンタのカウント値をインクリメントしておくことによって検出可能である。例えば、学習回数は、学習勾配や学習曲率と共にＩＤアドレスで指定されるメモリ領域上に記憶される。

Note that the number of times of learning here can be detected by providing the navigation ECU 4 with a counter indicating the number of times of learning, and incrementing the counter value every time an instantaneous gradient is obtained. For example, the number of learnings is stored in a memory area specified by an ID address together with a learning gradient and a learning curvature.

  The contents stored here are deleted when the amount stored with respect to the prescribed amount of the memory area reaches a predetermined ratio or when the remaining capacity reaches a predetermined amount. . That is, it is partially erased according to the storage amount or the storage allowable amount. The order of deletion at this time is determined according to at least one of the learning time, the degree of learning, the current location or a spatial relationship with a predetermined location (for example, home). For example, the date and time of learning are oldest or far from the current location. In order.

  The block configuration of the navigation ECU 4 and the engine ECU 6 described so far, that is, the configuration for realizing the function of calculating and storing the learning data in the navigation ECU 4 and the engine ECU 6, and various sensors used for the calculation of the learning data This corresponds to a travel load information learning system that implements a learning method.

  The navigation ECU 4 further includes a search distance determination unit 4d, a future trajectory information generation unit 4e, and a trajectory (road) shape calculation unit 4f.

  The search distance determination unit 4d determines the search distance of the range to be searched in the road map information, that is, how far the road map information from the current position of the vehicle is necessary. The search distance determination unit 4d sets, for example, several nodes (for example, 10 nodes) or several segments (for example, 11 segments) in the traveling direction from the current position as the search distance, or detection from the wheel speed sensors 1a to 1d. The search distance is set according to the vehicle speed calculated based on the signal. When the search distance is set according to the vehicle speed, for example, the search distance is set to be longer as the vehicle speed is faster.

  The future trajectory information generation unit 4e generates trajectory information that the vehicle will travel in the future. For example, the trajectory information for several nodes (for example, 10 nodes) in the traveling direction from the current position or the trajectory information that the vehicle will travel during a predetermined time (for example, several seconds) from the present is generated as future trajectory data. The Specifically, turning curvature data for several nodes ahead in the traveling direction, learning gradient data for several nodes ahead in the traveling direction, and distance data to several nodes ahead in the course direction are extracted as future trajectory data. The future trajectory data is generated by extracting learning contents stored in the map DB 4c in the navigation ECU 4 and road map information stored in a CD-ROM or DVD-ROM used in the navigation system. That is, the future trajectory data is obtained by searching and reading the learning gradient, learning curvature data, and distance data stored in the ID address corresponding to several nodes ahead in the course direction from the map DB 4c, CD-ROM, DVD-ROM, etc. Is generated.

  The track (road) shape calculation unit 4f, based on the track information generated by the future track information generation unit 4e, the future gradient vector that becomes the gradient information of the road surface where the vehicle will travel in the future, and the road surface A future curvature vector as curvature information is obtained. That is, based on the future trajectory data for several nodes located on the trajectory on which the vehicle already obtained will travel in the future, for example, the future slope of the road surface on which the vehicle will travel during a predetermined time from the present Find vectors and future curvature vectors. The distance that the vehicle will travel within a predetermined time from the present is calculated based on the vehicle speed, and the future gradient vector and the future curvature vector are calculated using the future trajectory data of the nodes that fall within that distance.

  For example, the future gradient vector is obtained if the distance to the future position where the vehicle will travel several seconds after the current position and the learning gradient stored for each node are known. Further, for example, the gradient to each future position based on the current position, the distance to the future position after the elapse of 2 seconds, the elapse of 4 seconds, and the future position after the elapse of 6 seconds, and the learning gradient of the current position and each future position. May be obtained as a time-series vector, or the maximum value or the average value thereof may be set as a future gradient vector.

  The future curvature vector can also be obtained by the same method as the future gradient vector based on the learning curvature at each of the current position and the future position.

  The future gradient vector and the future curvature vector obtained as described above are sent to the engine ECU 6 and used as parameters of various processes executed by the engine ECU 6.

  The engine ECU 6 also has a block configuration shown in FIG. As shown in this figure, the engine ECU 6 includes a driver request torque calculation unit 6c and a vibration suppression control unit 6d. The driver request torque calculator 6c receives the detection signal from the accelerator pedal sensor 2d, and sets the engine control amount so as to generate a drive torque according to the driver's drive request. The vibration suppression control unit 6d adjusts the engine control amount so as to suppress the vibration generated in the vehicle body as small as possible according to the turning state of the vehicle. As a vibration control method, for example, one disclosed in Japanese Patent Application Laid-Open No. 2000-233668 can be employed. The vibration control unit 6d receives the future curvature vector obtained by the navigation ECU 4, and predicts the turning state of the vehicle based on the future curvature vector, so that the engine control amount can be adjusted according to the prediction. .

  In the case where the future curvature vector is not yet obtained, that is, the road on which the vehicle is currently traveling for the first time, the ID address corresponding to the node or segment on the road in the map DB 4c of the navigation ECU 4 is used. If the instantaneous curvature has not yet been stored, the instantaneous curvature obtained for the first time can be used instead of the future curvature vector.

  In this way, the engine ECU 6 determines the engine control amount, outputs a control signal so as to finally obtain the required engine control amount, and adjusts the engine speed and the throttle opening, thereby reducing the required torque. It has come to be realized.

  The engine ECU 6 further has a block configuration shown in FIG. As shown in this figure, the engine ECU 6 includes a gradient load estimator 6e, a turning load estimator 6f, a rolling / air resistance estimator 6g, and a travel environment load fluctuation calculator 6h, and performs load fluctuations according to the vehicle environment. It has the required function.

  The gradient load estimating unit 6e estimates a gradient load that becomes a load corresponding to the gradient of the road surface. This gradient load is estimated based on the future gradient vector obtained by the navigation ECU 4, and it is estimated that the gradient load is larger as the future gradient vector is larger. Also, if the future gradient vector indicates an uphill, it acts as a load on the drive system, so the gradient load is indicated by a positive value, and if it indicates a downhill, it is a negative value. Indicated by In the case where the future gradient vector is not yet obtained, that is, the road on which the vehicle is currently traveling for the first time, the map DB 4c of the navigation ECU 4 has an ID address corresponding to the node or segment on that road. If the instantaneous gradient is not yet stored, the instantaneous gradient obtained for the first time can be used instead of the future gradient vector.

  The turning load estimating unit 6f estimates a turning load that becomes a load corresponding to the turning of the vehicle. The estimation of the turning load is executed based on the future curvature vector obtained by the navigation ECU 4, and it is estimated that the turning load increases as the future curvature vector increases. Even in this case, if the future curvature vector has not yet been obtained, the instantaneous curvature obtained for the first time can be used in place of the future curvature vector.

  The rolling / air resistance estimation unit 6g estimates the rolling resistance of each wheel and the air resistance to the vehicle body. Information relating to the vehicle speed obtained by the brake ECU 9 based on the detected values of the wheel speed sensors 1a to 1d is input to the rolling / air resistance estimating unit 6g. Based on the vehicle speed obtained from this information. The rolling resistance and air resistance are estimated, and it is estimated that each of these resistances increases as the vehicle speed increases.

  The traveling environment load fluctuation calculating unit 6h calculates how much the load of the driving system has on both the positive and negative sides with respect to the traveling environment by adding the turning load and rolling resistance / air resistance from the gradient load.

  The T / M-ECU 8 has a block configuration shown in FIG. As shown in this figure, the engine ECU 6 includes a driver request shift stage calculator 6i, a drive torque required gear stage calculator 6j, a load torque required gear stage calculator 6k, a turning stabilization gear stage calculator 6m, a shift constraint calculation. Unit 6n and first to third gear stage arbitration units 6o, 6p, 6q, which have the function of setting an optimum gear stage in T / M 7 and realizing a shift according to the setting.

  The driver request shift stage calculation unit 6i detects a gear stage according to a direct operation of the driver based on a signal from the shift position sensor 3b. For example, the shift of the automatic transmission can select each range of D (driving), N (neutral), R (rear), and P (parking) in which 1st, 2nd, and 1st to 4th are automatically shifted. In some cases, when a range other than D is selected, the gear stage that matches the selected range remains as it is for the driver to operate, regardless of the gear arbitration contents of the following elements 6j to 6p. The gear stage is set accordingly.

  The drive torque required gear stage detection calculation unit 6j detects a gear stage that satisfies the drive torque obtained by the driver request torque calculation unit 6c in the engine ECU 6. The detection of the gear stage according to the driving torque is the same as that conventionally performed.

  The load torque required gear stage calculation unit 6k detects a gear stage that satisfies the torque range necessary to compensate for the load fluctuation obtained by the traveling environment load fluctuation calculation unit 6h. Specifically, it is sufficient that the drive torque capable of compensating for the load fluctuation is generated by the gear stage detected by the drive torque required gear stage calculation unit 6j, but this may not be sufficient. In such a case, a gear stage that can generate a larger driving torque is selected by the load torque required gear stage calculation unit 6k. For example, even when the gear stage detected by the drive torque required gear stage calculation unit 6j is the fourth speed, the third speed may be detected by the load torque required gear stage calculation unit 6k.

  The turning stabilization gear stage calculation unit 6m receives the future curvature vector obtained by the navigation ECU 4, and when the vehicle makes a turn corresponding to the future curvature vector, the driver performs an accelerator operation to operate the vehicle. This is to detect the gear stage required to generate Unlike the gear stage that satisfies the driving torque, the turning stabilization gear stage calculation unit 6m calculates a gear stage that generates a driving torque that enables the driver to correct the trajectory by the accelerator work during turning. Specifically, there are situations where the accelerator is slowed down when entering the corner at the overspeed limit, the speed is lowered and the trajectory is corrected to the inside, and the vehicle is stabilized and accelerated at the exit of the corner. Under such circumstances, it is expected that a negative driving torque (engine brake) will work moderately when the driver's accelerator is off, and that driving will be applied without excessive downshifting when the accelerator is on. For this reason, even when the gear stage detected by the drive torque required gear stage calculation unit 6j is the fourth speed, the turning stabilization gear stage calculation unit 6k may detect the third speed. Even in this case, if the future curvature vector has not yet been obtained, the instantaneous curvature obtained for the first time can be used in place of the future curvature vector.

  The shift restriction unit 6n determines the gear position of T / M7 from the viewpoint of engine protection and temperature environment (early activation of cooling water, engine oil, catalyst, etc.). For example, even if the gear stage selected at the current vehicle speed is the second speed, if it is not preferable to actually shift to the second gear stage, the engine speed will be too high. Therefore, there is a restriction that the gear stage becomes 3rd speed or higher. Conversely, when the engine speed is too low, or when the cooling water temperature is low, there is a restriction that the first speed is set.

  The first gear stage arbitration unit 6o mediates detection results of the drive torque required gear stage calculation unit 6j and the load torque required gear stage 6k. The first gear stage arbitration unit 6o selects a low-speed gear stage that can obtain a larger driving torque from the detection results of the driving torque required gear stage calculation unit 6j and the load torque required gear stage calculation unit 6k.

  The second gear stage arbitration unit 6p mediates the arbitration result of the first gear stage arbitration unit 6o and the detection result of the turning stabilization gear stage calculation unit 6m. The second gear stage arbitration unit 6p selects a low-speed gear stage that can obtain a larger driving torque among the detection results of the first gear stage arbitration unit 6o and the turning stabilization gear stage calculation unit 6m. For example, the second gear stage arbitration unit 6p restricts the signal from being shifted up when a signal to shift up to a gear position higher than the detection value of the turning stabilization gear stage calculation unit 6m comes from the first gear stage arbitration unit 6o. impose. If the curvature information used is not an instantaneous curvature but a future curvature vector, it can be shifted down.

  The third gear stage arbitration unit 6q mediates the arbitration result of the second gear stage arbitration unit 6p and the determination results of the driver request shift stage calculation unit 6i and the shift constraint unit 6n. Specifically, the third gear stage arbitration unit 6q gives priority to the determination results of the driver request shift stage calculation unit 6i and the shift constraint unit 6n over the arbitration result of the second gear stage arbitration unit 6p. A selection is made. That is, when the shift requested by the driver is in a range other than D, the gear stages corresponding to those ranges are preferentially selected, and in the case of the D range, the second gear stage arbitration unit 6p. The gear stage corresponding to the arbitration result is selected. Then, the gear stage is finally determined within the range of the gear stage determined by the shift restricting unit 6n, and engine protection or the like is performed. For example, when the gear stage according to the arbitration result of the second gear stage arbitration unit 6p is a lower speed gear than the gear stage determined by the shift regulation unit 6n, the determination result of the shift regulation unit 6n The gear stage corresponding to is preferentially selected.

  In this way, a control signal for realizing a shift to achieve the selected gear stage is output from the T / M-ECU 8, and a gear shift of T / M7 is performed.

  In the vehicle driving force control system having such a configuration, the learning gradient and the learning curvature are stored in the map DB 4c in the navigation ECU 4 each time the vehicle travels. That is, the basic data of the road running load information is not stored in advance on a CD-ROM, DVD-ROM or the like, but the road running load information serving as basic data is learned. Therefore, it becomes possible to learn the road gradient and curvature corresponding to the modification of the road shape and gradient, etc., and it is possible to cope with the change without buying a CD-ROM or DVD-ROM each time. It becomes possible. Further, by storing basic data in advance, it is possible to prevent an increase in the amount of data memory such as a CD-ROM or DVD-ROM.

  Then, based on such learning gradient and learning curvature, it is possible to pre-select the optimum gear stage of T / M7 at each location on each road based on various calculations in the engine ECU 6 and the T / M-ECU 8. Become. For this reason, the following effects can be obtained on curved roads and uphill / downhill roads.

  When approaching a curved road, the driver also performs a brake pedal operation. However, the driver tries to correct the vehicle speed and the trajectory in response to the change in the longitudinal speed and the lateral direction on the curved road mainly by the accelerator pedal operation. In such a case, if the gear speed is such that the engine brake is difficult to apply or only a low torque can be obtained, such as the fourth speed, for example, the correction suitable for the intention cannot be made only by operating the accelerator pedal.

  Accordingly, as shown in the present embodiment, the turning stabilization gear stage calculation unit uses the learning curvature to detect in advance the gear stage necessary to correspond to the curved road and before reaching the curved road. By shifting the gear to the corresponding gear stage at this stage, it is possible to generate engine brake and acceleration torque that can appropriately cope with the above change on the curve road only by operating the accelerator pedal. For this reason, the driver can concentrate on the accelerator pedal operation and can travel on a curved road with peace of mind.

  In the current automatic transmission, when the brake is depressed, the automatic transmission may shift to a gear stage that generates an engine brake greater than the required braking force depending on the depressed state. According to the vehicle driving force control system of the present embodiment, it is possible to provide a more comfortable vehicle travel to the driver because such an unnecessary engine brake is not generated.

  In addition, when you reach an uphill road, you depress the driver's accelerator pedal so that the vehicle can climb the uphill road, but an accelerator pedal operation is performed and a time lag occurs until the gear is downshifted. The driving force that can be dealt with cannot be generated immediately.

  Therefore, as shown in the present embodiment, the load torque required gear stage calculation unit uses the learning gradient to detect a gear stage necessary to correspond to the uphill road in advance and before reaching the uphill road. By shifting to the corresponding gear stage at this stage, it is possible to climb the uphill road more smoothly.

  On the other hand, the downhill road is the same as the uphill road, and the accelerator pedal is stopped so that the vehicle speed does not increase too much on the downhill road, and the engine brake is generated. However, a time lag occurs until the engine brake is generated. Moreover, even if engine braking is generated, sufficient engine braking may not be obtained.

  Therefore, as shown in the present embodiment, the load torque required gear stage calculation unit 6k uses the learning gradient to detect in advance the gear stage necessary to cope with the downhill road, and approaches the downhill road. The vehicle speed can be controlled more accurately by shifting to the corresponding gear stage in the previous stage.

  As described above, by using the vehicle driving force control system shown in the present embodiment, it is possible to realize a shift corresponding to a curved road or an uphill / downhill road. And since the response to the curve road and the uphill / downhill road is performed by the shift of such an automatic transmission, the area that can be operated by the will of the driver is reduced compared to the case of correcting the engine control amount. In other words, the optimum driving force range can be realized without saturating the driving force against the will of the driver.

  However, as described above, in the case where the learning gradient and the learning curvature are not obtained, that is, when traveling for the first time on a road surface that has not yet been traveled, the instantaneous gradient and the instantaneous curvature obtained at that time are used. Yes. Therefore, in such a case, it is impossible to prepare for a curved road or an uphill / downhill road in advance, but it is possible to perform a shift according to at least the instantaneous gradient and the instantaneous curvature obtained at that time.

  In the driving force control device disclosed in Patent Document 1, vehicle speed information and driving state associated with road map information are stored at each location of each road. For this reason, these vehicle speed information and driving state can be used only for engine control amount correction, and versatility is poor. However, in the vehicle driving force control system in the present embodiment, only the gradient and curvature are stored in the map DB 4c. These gradients and curvatures are physical quantities of a trajectory on which the vehicle travels, and are physical quantities that exert a disturbing effect on the motion of the vehicle, such as a load on driving. It also has the meaning of which direction the vehicle should go. Therefore, it can be used for suspension control, steer control, headlight light distribution control, etc. in addition to the shift adjustment of T / M7, and can be said to be rich in versatility. It is possible to use these gradients and curvatures as calculation parameters applied to various processes, specifically, other than T / M7 speed adjustment.

(Second Embodiment)
In the present embodiment, the learning data is subjected to weighting with respect to the number of times of learning in the first embodiment, and the gradient and curvature at each node position are obtained. Since other parts are the same as those in the first embodiment, description thereof is omitted here.

  Since the instantaneous gradient is calculated every time the vehicle travels on the road, it is already obtained on the road surface on which the vehicle has traveled before and is stored as learning data. In this embodiment, using this learning data, the learning gradient is weighted to obtain the instantaneous gradient. The details of the weighting process for the number of learnings will be described below with reference to FIG.

  Fig. 9 shows the relationship between the distance and gradient for each segment. The figure above the page shows the nodes and segments in a plane on the map, and the figure below the page shows them on the map. It is shown in cross section. In this figure, an arbitrary segment is represented as sn, a distance between the segments sn is represented as dn, a gradient formed with respect to a horizontal plane is represented as gn, and a relative gradient between adjacent segments is represented as gn ′.

For example, in the case where the learning gradient is stored in the map DB 4c in units of segments, when the learning gradient is converted into gradient data in node units and used as a learning gradient, the number of learning times of each of two segments adjacent to an arbitrary node is n. _Assuming that _n and n _{n + 1} are expressed, the gradient data gn _n averaged from the number of learning times is expressed by the following mathematical formula.

この数式は、学習回数が多いほど、信頼性の高い勾配データであるということを前提として、学習回数に応じて重み付け平均を行ったものである。このように求められる勾配データｇｎ_nが各ノードの学習勾配として読み出される。

This formula is a weighted average according to the number of learnings on the premise that the gradient data has higher reliability as the number of times of learning increases. The gradient data gn _n obtained in this way is read out as the learning gradient of each node.

  The number of learning can be detected by providing the navigation ECU 4 with a counter indicating the number of learning, and incrementing the counter value every time an instantaneous gradient is obtained. For example, the number of learning is counted by a counter provided on a memory specified by an ID address together with a learning gradient and a learning curvature, and is incremented when traveling in the corresponding section.

  As described above, the learning gradient at the node position can be obtained by performing weighting according to the number of times of learning data learning. In this way, it is possible to obtain learning data with higher reliability. Although the learning gradient has been described here, the learning curvature is also weighted by the same method as the learning gradient.

(Third embodiment)
In the present embodiment, the segment distance is weighted with respect to the first embodiment, and the gradient and curvature at each node position are obtained. Since other parts are the same as those in the first embodiment, description thereof is omitted here.

  When a learning gradient corresponding to a segment with a long distance is compared with a learning gradient corresponding to a segment with a short distance, it is assumed that the former is more reliable than the latter as indicating the gradient. Therefore, in this embodiment, weighting according to the segment distance is performed. The details of the weighting process according to the segment distance will be described below based on FIG. 10 shown in the second embodiment.

For example, in the case where the learning gradient is stored in the map DB 4c in units of segments, when the learning gradient is converted into gradient data in node units to obtain the learning gradient, the distance between each of two segments adjacent to an arbitrary node is dn, When represented by dn + 1, the slope data gd _n averaged from the distance of the segment is expressed by the following equation.

この数式を用いることで、セグメントの距離に応じ、その距離が長いほど重みが高い演算が成される。このように求められる勾配データｇｎ_nが各ノードの学習勾配として読み出される。

By using this mathematical formula, an operation with a higher weight is performed as the distance becomes longer according to the distance of the segment. The gradient data gn _n obtained in this way is read out as the learning gradient of each node.

  Note that the data regarding the distance of the segment can be calculated from the position coordinate string data recorded in the map DB 4c and the connection information between the two nodes forming the segment.

  As described above, the learning gradient can be obtained by performing weighting according to the segment distance. In this way, it is possible to obtain learning data with higher reliability. Although the learning gradient has been described here, the learning curvature can be weighted by the same method as the learning gradient, and learning data with higher reliability can be obtained.

(Other embodiments)
(1) In each of the above embodiments, the turning curvature of the vehicle is learned and stored as the learning curvature. However, the turning curvature is not stored and can be obtained by calculation each time. is there. For example, if you select three adjacent nodes on the road on which the vehicle will travel and find an arc that passes through all three points, the radius of the arc becomes the radius of curvature. It can be obtained.

  Of course, the turning curvature is basically memorized, and the turning is performed by the above calculation only in the case where learning is not possible due to the occurrence of a defect in the calculation of the curvature radius or in the initial state. It is also possible to obtain the curvature.

  For example, in a CD-ROM or DVD-ROM of a navigation system, the absolute angle with respect to the ground axis is set for one segment formed between two nodes from the position coordinates stored for each node. By calculating, a relative angle formed between two adjacent segments can be calculated. This angle is represented as an angle θn formed by each segment sn with respect to the reference vector, as shown in the vector diagram of FIG.

  Further, assuming that the angles of the segments sn stored in two adjacent nodes are θn and θn + 1, respectively, the relative angle between the segments (hereinafter referred to as the relative segment angle) θn ′ is θn ′ = θn−θn + 1. It is represented by If this relative segment angle θn ′ takes a positive value, it becomes a right curve, and if it takes a negative value, it becomes a left curve.

  Then, as shown in the schematic diagram of FIG. 12, assuming that the radius of the arc passing through each node and the three nodes before and after the node is rn and the distance between the nodes is dn, the radius rn of the arc is the relative segment angle θn. 'And the distance dn between the nodes are expressed by the following equation.

このようにして曲率半径となる半径ｒｎが求められるので、曲率半径の逆数の式（数４の逆数）を用いることで演算上のゼロ割りを起こすことなくカーブの形状を求めることが可能である。

Since the radius rn as the radius of curvature is obtained in this way, it is possible to obtain the shape of the curve without causing an arithmetic division by zero by using the expression of the reciprocal of the radius of curvature (the reciprocal of Equation 4). .

  (2) In the second and third embodiments, an example in which different weighting processes are performed separately is shown, but both weighting processes can be performed together. For example, it is possible to take the average of the results of the weighting process according to the number of learnings and the weighting process according to the segment distance. In this case, the gradient data gn 'is expressed by the following equation.

また、学習回数およびセグメントの距離を同時に考慮して勾配データを求めることも可能である。この場合、勾配データｇｎ’は、次式で示される。

It is also possible to obtain gradient data by simultaneously considering the number of learnings and the segment distance. In this case, the gradient data gn ′ is expressed by the following equation.

これらに示されるように、上記第２、第３実施形態を組み合わせて学習データを求めることが可能であり、より信頼性の高い学習データとすることができる。

As shown in these, learning data can be obtained by combining the second and third embodiments, and learning data with higher reliability can be obtained.

  (3) In the second embodiment, the learning frequency is obtained from the count value of the counter corresponding to each node or segment, but the learning data is stored only in either the segment unit or the node unit, When the learning number is not counted, the other learning number can be obtained based on the learning number of the stored learning data.

  For example, when learning data is stored in segment units, it is converted into learning frequency data in node units and output. That is, if the learning counts of two segments adjacent to an arbitrary node are nn and nn + 1, and the distances are dn and dn + 1, the learning counts of the nodes are averaged from the segment distances, and are expressed by the following equation: Will be.

（４）上記各実施形態では、路面の勾配推定の方法の一例として、特開平８−５３０２４号公報に示される方法を適用することを説明したが、この他の方法に基づいても路面の勾配推定を行うことができる。

(4) In each of the above embodiments, the application of the method disclosed in Japanese Patent Application Laid-Open No. 8-53024 has been described as an example of the method of estimating the road surface gradient. Estimation can be performed.

  For example, an acceleration sensor is provided in the vehicle, and the acceleration gx of the vehicle is detected from a detection value obtained by the acceleration sensor. On the other hand, after obtaining the wheel speed Vw based on the detection signals from the wheel speed sensors 1a to 1d, the vehicle speed V is obtained by a known method such as selection thereof, and the differential value ΔV is obtained. The difference between the acceleration gx and the differential value ΔV is obtained.

  The acceleration gx obtained from the detection value of the acceleration sensor includes not only a pure acceleration of the vehicle but also an acceleration component based on gravity. On the other hand, the vehicle speed differential value ΔV obtained based on the detected values of the wheel speed sensors 1a to 1d indicates the pure acceleration of the vehicle. Therefore, these differences correspond to g sin θ (g: gravity) when assuming a gradient when the road surface inclination angle is θ, and the following mathematical formula is established.

(Equation 8)
(Gx−ΔV) = gsinθ
Since the gradient corresponds to tan θ, tan θ corresponding to the gradient can be obtained by obtaining θ from the relational expression shown in Equation 8.

  In addition, if a system that communicates with a service center via a communication device and a communication network that are being developed in recent years, that is, a communication device and a communication network provided in the vicinity of a road, is used, It is also possible to obtain slope information on roads that have no road and more detailed slope information on roads that have traveled. Of course, using this method also makes it possible to acquire travel load information relating to other vehicles.

  Furthermore, it is also possible to measure the road altitude with GPS signals received by the navigation system and obtain the road gradient from the altitude difference.

  (5) It is also possible to perform gradient estimation with higher accuracy by performing error correction on the above-described conventional road surface gradient estimation method. In this way, it is possible to execute vehicle driving force control based on a more accurate gradient.

  Examples of the conventional method for estimating the road surface gradient described above include the following.

  One is to compare the acceleration (hereinafter referred to as the detection value) obtained from the detection signal of the longitudinal acceleration sensor and the vehicle acceleration obtained from the change in the vehicle speed. For example, the detection value of the longitudinal acceleration sensor is determined from these differences. The gradient of the road surface is estimated by extracting the gravitational acceleration component included in the road.

  The other is a comparison of the acceleration estimated from the driving force of the vehicle and the vehicle acceleration determined from the change in vehicle speed. The acceleration estimated from the driving force and the actual vehicle acceleration are affected by the gradient of the road surface. Since a difference occurs, the road gradient can be estimated from this difference.

  However, in these methods, the centrifugal force and the side slip angle have an effect on the detection signal of the acceleration sensor when the vehicle is turning, or the drag (the action that occurs in the direction opposite to the moving direction of the vehicle body due to the tire being snaked during the steering operation). It is desirable to remove the error because it includes an error due to the force being a driving load. Therefore, it is preferable to remove the error as follows.

  First, an error removal method for the first gradient estimation method will be described with reference to a schematic diagram shown in FIG.

  FIG. 13A shows a case where the vehicle is traveling on a curved road at a traveling direction speed V. FIG. Under the situation shown in this figure, if the detected value of the acceleration sensor is Gxs, the centrifugal force is Gy, the skid angle is β, and the detected value of the vehicle speed is Vxs, the detected value Gxs of the acceleration sensor is as shown in FIG. As a vector shown in the schematic diagram, the detected value Vxs of the vehicle speed is expressed as a vector shown in the schematic diagram of FIG. Therefore, it is expressed as the following equation.

(Equation 9)
Gxs = Gxcosβ-Gysinβ
(Equation 10)
Vxs = Vcosβ
Further, when the equation 10 is differentiated, it is expressed as follows.

(Equation 11)
Vxs ′ = V′cosβ−Vβ′sinβ
The road surface gradient θ currently being traveled is expressed by the following equation (12), and the above equations (9) and (11) are converted into the following equations (13) and (14). By substituting 14, Equation 15 is derived, and an error due to the influence of the centrifugal force Gy and the side slip angle β is expressed as a term of Gysinβ−Vβ′sinβ expressed by Equation 15 from the detection value of the acceleration sensor.

(Equation 12)
θ = asin {(Gx−V ′) / 9.8}
(Equation 13)
Gx = (Gxs + Gysinβ) / cosβ
(Equation 14)
V ′ = (Vxs ′ + Vβ′sinβ) / cosβ
(Equation 15)
θ = asin {(Gxs−Vxs ′ + Gysinβ−Vβ′sinβ}
/(9.8cosβ)
Therefore, for example, if the centrifugal force Gy and the side slip angle β are estimated from the steering wheel angle θs, the vehicle speed Vxs, etc., the term indicating the error shown in the above equation 15 can be obtained, and this error is removed and the gradient θ Can be obtained. Thereby, the accuracy of gradient estimation can be improved.

  As another gradient estimation method, the gradient is obtained with high accuracy by subtracting the drag and air resistance, which are rolling resistance and vehicle deceleration force components generated during turning, from the estimated driving force. The drag acting in the longitudinal direction of the vehicle body can be calculated from the forces acting on the tire and road surface using a known two-wheel vehicle model, which is given as a function of the front wheel steering angle, rear wheel steering angle, and vehicle speed. It is done. Air resistance is given as a function of vehicle speed. By subtracting the drag and air resistance of the front and rear wheels thus obtained from the estimated driving force, the load when traveling on a gradient road surface excluding the effects of rolling resistance and turning of the tire is obtained. Therefore, by dividing this load by the vehicle weight M, it can be converted into an acceleration component. If this value is used as Gxd in place of Gxs in Equation 15 in the first gradient estimation method described above, the gradient θ can be obtained by removing the error due to the drag. This also improves the accuracy of gradient estimation.

  Further, during acceleration / deceleration, the gravitational acceleration component is superimposed on the detection signal of the acceleration sensor due to the influence of pitching vibration. For this reason, it is preferable to remove this influence. For example, as shown in FIG. 14, when it is assumed that the pitch angle with respect to the horizontal line of the vehicle body in the vehicle is θp, the pitch angle θp is expressed as a function of the differential value Vxs ′ of the vehicle speed detection value Vxs. If this is expressed, the detected value Gxs of the acceleration sensor is expressed as shown in Equation 17, and when this equation is converted, the acceleration Gx in the traveling direction parallel to the road surface is expressed as shown in Equation 18.

(Equation 16)
θp = f (Vxs ′)
(Equation 17)
Gxs = Ggsinθp + Gxcosθp
= Ggsin {f (Vxs ')} + Gxcos {f (Vxs')}
(Equation 18)
Gx = [Gxs−Ggsin {f (Vxs ′)}] / cos {f (Vxs ′)}
In this way, the acceleration Gx in the traveling direction parallel to the road surface from which the influence of the pitching vibration has been removed is obtained, and by using this to perform the gradient estimation, the accuracy of the gradient estimation can be further improved.

  Although two methods for estimating the road surface gradient have been described here, it is also possible to use a combination of these two methods.

  For example, the acceleration obtained from the detection value of the acceleration sensor is likely to cause an offset error depending on the sensor mounting angle and temperature characteristics. Further, the acceleration estimated from the driving force is likely to cause an error because a large error occurs due to slipping of the torque converter, or the braking force becomes a load on the driving source during the braking operation.

  For this reason, the value when the difference in acceleration used for each method passes the low-pass filter, that is, the offset amount is obtained, and the value obtained by subtracting the offset amount from the detection value by the acceleration sensor is used as the acceleration from which the error is removed. It is also possible to estimate the road gradient.

  That is, when the acceleration estimated from the driving force is Gxd and the acceleration obtained from the detection value of the acceleration sensor is Gxs, the offset amount is expressed by the following equation.

(Equation 19)
Offset amount = LPF (Gxs−Gxd)
However, LPF represents a low-pass filter, and the time constant in the LPF is set on the order of seconds. The calculation using this mathematical expression is performed when conditions such as not braking, not extremely low speed, and the gear is not in the neutral (N) or parking (P) range are obtained. A value obtained by subtracting the offset amount from the acceleration Gxs obtained from the detection value of the acceleration sensor (Gxs−offset amount) can be used for estimating the road surface gradient. As a result, it is possible to perform road surface gradient estimation based on acceleration with the effect of offset error suppressed, and to further improve gradient estimation accuracy.

  When it is considered that the braking force is affected by the braking operation, the value obtained by subtracting the offset amount from the acceleration Gxd estimated from the driving force described above is used as the acceleration. The gradient estimation may be performed using the acceleration Gxd estimated from the driving force. Whether or not it is during such a brake operation can be determined by, for example, a detection signal from a brake pedal operation amount sensor (not shown) input to the brake ECU 9.

  (6) In each of the above embodiments, the road surface gradient and the curvature are learned as the vehicle running load information. However, it is also possible to learn the slope of the road surface in the left-right direction.

  (7) In each of the above embodiments, it is only necessary to obtain learning data in all of the nodes or all of the segments. However, for example, there may be a case where learning data at a specific node or segment cannot be obtained due to a high vehicle speed. . In such a case, it is also possible to interpolate the learning data of a specific node or segment based on the learning data obtained at the preceding node or segment.

  (8) In the above embodiment, weighted addition is performed in consideration of cases where accurate traveling load information cannot be obtained, such as when the vehicle is extremely low speed or when slipping occurs. For example, in the case of the above-described embodiment, when accurate traveling load information cannot be obtained, addition with a weight of zero is performed by removing it from the accumulated value. Such weighted addition is information that conveys to the navigation ECU 4 data related to the instantaneous gradient and the like while performing at least one of the section distance for which the traveling load information is estimated, the reliability of the estimated information, the learning degree, and the gradient estimation. This can be performed according to the state of the engine ECU 6 serving as a transmission source.

(9) Notification of branching In the embodiments described so far, one road that progresses in the future is selected, and load information related to the future trajectory is read and used for control. In many cases, there are branches on the track, and the road traveling in the future is not always the selected track, only the driver knows. If the trajectory is selected incorrectly, control is performed based on incorrect load information, which is not preferable. Therefore, when transmitting load information to the control side as a time series vector etc., the existence of a branch is also in the form of a time series vector so that it can be understood how far the branch exists and where it is after the branch. Good to tell.

  Also, the number of times of learning may be conveyed by a vector that indicates in which section learning is possible. These may be flags that information after branching or unlearned information is included in the transmission information section. And the control side can avoid control by erroneous information by switching to use the instantaneous value based on these vectors or flags.

  In this way, branching and unlearned information are also output from the navigation ECU 4 so that the vehicle control using the traveling load information, for example, the above-described vehicle driving force control system, branches. It is possible to execute vehicle control corresponding to unlearning.

  Note that detection of whether or not there is a branch on the road on which the vehicle is traveling is performed based on road map information stored in a CD-ROM or DVD-ROM of the navigation system, and learning of load information is performed. Detection of a section that has been made (or a section that has not been learned) is performed based on the contents stored in the storage unit of the navigation ECU 4. Therefore, the navigation ECU 4 that can extract the road map information and the contents stored in the storage unit constitutes a branch detection means and an estimated section detection means.

(10) About the learning mechanism Although this plan showed the average by the moving distance, the number of learnings, and the weighting by the segment distance as the load information storing / reading algorithms, the navigation ECU 4 in the electronic control system of the entire vehicle It has a feature that can associate control. Therefore, when there is a request for learning of other control parameters as well as the load information, it can be used as an algorithm for learning and reading out the parameters.

(11) Storage means The storage medium used in the storage means may be a flash memory or a hard disk. If a CD or DVD storing road map information is writable, it is stored in the predetermined memory area. May be.

It is a figure which shows schematic structure of the driving force control system for vehicles in 1st Embodiment of this invention. It is the figure which showed a part of block configuration of navigation ECU4 in the vehicle driving force control system shown in FIG. It is the figure which showed a part of block configuration of engine ECU6 in the driving force control system for vehicles shown in FIG. It is the figure which showed a part of block configuration of navigation ECU4 in the vehicle driving force control system shown in FIG. It is the schematic diagram which showed the relationship between the content memorize | stored in CD-ROM, DVD-ROM, etc. which are used for a navigation system, and the memory content to the memory | storage part with which navigation ECU4 was equipped. It is the figure which showed a part of block configuration of engine ECU6 in the driving force control system for vehicles shown in FIG. It is the figure which showed a part of block configuration of engine ECU6 in the driving force control system for vehicles shown in FIG. It is the figure which showed a part of block configuration of engine ECU6 in the driving force control system for vehicles shown in FIG. It is the figure which showed the track | orbit of a node and a segment planarly and in cross section. It is the figure which showed the relationship between each segment and its distance. It is the figure which showed the relation of the relative angle of the segments which adjoin each node. It is a figure for demonstrating how to obtain | require the radius of the circular arc which passes three nodes. (A) is the model which showed the case where the vehicle is drive | working the curve road at the advancing direction speed V, (b), (c) is in the condition where the vehicle is drive | working like (a). It is the schematic diagram which showed each parameter. It is a schematic diagram which shows the influence of the pitching vibration at the time of acceleration.

Explanation of symbols

1a to 1d: wheel speed sensor, 2a: engine speed sensor,
2b: throttle opening sensor, 2c: intake air temperature sensor,
2d ... accelerator pedal sensor, 3a ... yaw rate sensor,
3b: shift position sensor, 4 ... navigation ECU, 5 ... engine,
6 ... Engine ECU, 7 ... T / M, 8 ... T / M-ECU, 9 ... Brake ECU.

Claims (14)

Current position detecting means (4, 4c) for obtaining a current position where the vehicle is traveling based on road map information when the vehicle is traveling on the road;
Traveling load information estimating means (4, 4a, 4b, 6, 6a, 6b) for estimating traveling load information of the road at the current position;
A traveling load information learning system comprising storage means (4, 4c) for learning and storing the traveling load information in association with the current position. The travel load information learning system according to claim 1, wherein the travel load information estimation means estimates at least one of a gradient and a curvature of the road as the travel load information. The storage means collects the travel load information estimated by the travel load information estimation means, obtains a value to be stored according to at least one of a travel amount in a predetermined area, a reliability of the estimated information, and a learning degree, The travel load information learning system according to claim 1, wherein the travel load information learning system is stored as learning content of the travel load information. The storage means stores the traveling load information together with ID data indicating a predetermined space area corresponding to the current position,
The travel load information estimation means collects instantaneous information when traveling in the predetermined space area, and the travel load information to be stored in the storage means is an average value in the size of the predetermined space area The travel load information learning system according to claim 3, wherein the travel load information learning system according to claim 3 is estimated. The storage means, when collecting the instantaneous information, if the vehicle corresponds to at least one of when the vehicle is at a very low speed, when a slip occurs or when there is a malfunction in the traveling load information estimation means, 5. The travel load information learning system according to claim 4, wherein the average value is obtained by removing the instantaneous information in that case. The storage means stores the number of learning times of the travel load information, and from the travel load information estimated this time by the travel load information estimation means and the travel load information previously stored in the storage means, The travel load information learning according to any one of claims 3 to 5, wherein an average value of the travel load information corresponding to the number of times of learning is obtained, and the average value is stored as new travel load information. system. The travel load information learning system according to any one of claims 1 to 6, further comprising output means (4f) for outputting data relating to the travel load information stored in the storage means. A branch detection means (4) for determining a branch of a road on which the vehicle is traveling based on the road map information;
8. The travel load information learning system according to claim 7, wherein the output means outputs information related to the branch detected by the branch detection means together with the travel load information. Based on the road map information, comprising estimated section detecting means (4) for detecting a section in which the travel load information is estimated;
The travel according to claim 7 or 8, wherein the output means outputs information on a place where the travel load information detected by the estimation section detection means is not estimated together with the travel load information. Load information learning system. The travel load information stored in the storage means is partially erased according to the storage amount or storage allowable amount of the storage means, and at least one of the learning time, the learning degree, the current location or the spatial relationship with the predetermined location. The travel load information learning system according to any one of claims 1 to 9, wherein the one to be erased is determined according to the situation. The storage means comprises a structure in which the travel load information is associated with a travel point, and is divided into at least management information and travel environment information, the learning degree is registered as management information, and the load information is registered as travel environment information. The travel load information learning system according to any one of claims 1 to 10. A vehicle driving force control system comprising the learning load information learning system according to any one of claims 1 to 11,
Based on the travel load information and the road map information stored in the storage means, the travel load of the road on which the vehicle will travel is predicted, and the driving force of the vehicle is controlled according to the prediction result. A vehicle driving force control system comprising a driving force control unit (6). The vehicle driving force control system according to claim 12, wherein the driving force control unit adjusts a speed ratio in a transmission (7) of the vehicle according to the prediction result. When the vehicle travels on the road, the current position where the vehicle is traveling is obtained based on road map information, and the road load information of the road at the obtained current position is estimated, and then the travel is related to the current position. A road load information learning method characterized by learning and storing load information.

Images